Of Skyrails and Skytrains - Elevated rail in the Australasian urban transport environment - Elevated rail in the Australasian ...
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Australasian Transport Research Forum 2016 Proceedings
16 – 18 November 2016, Melbourne, Australia
Publication website: http://www.atrf.info
Of Skyrails and Skytrains - Elevated rail in the
Australasian urban transport environment
Ian Woodcock1, Scott Martin2
1
RMIT University, Melbourne, Victoria
2
University of Canberra, Bruce, ACT
Email for correspondence: ian.woodcock@rmit.edu.au & scott.martin@canberra.edu.au
Abstract
In recent years, elevated rail corridors have regained favour as a way to retrofit public
transport infrastructure into mature urban environments. Elevated rights-of-way are viewed
as particularly cost-effective solutions to high capital costs inherent in trenched and
tunnelled rail alignments. While elevated rights-of-way are extensively utilised for urban
freeway corridors in Australia and New Zealand, they are less utilised for urban rail.
Recent urban rail proposals by Australian state governments using elevated alignments
have met with public hostility from concerns about visual and audible impacts upon the
adjacent urban environments. The authors argue that elevated rail provides cost-effective
urban public transport corridors capable of delivering high-quality built environments,
especially around stations. Furthermore, elevated rail corridors, if done well can preserve
and enhance overall urban amenity for the public in comparison to trenched or tunnelled
alternatives.
This paper investigates claims of elevated rail's lower capital costs compared to trenched
and tunnelled alternatives using evidence from benchmarked costings from over 20 years of
Australian and international transport projects. Beyond construction costs, a range of other
costs and benefits associated with the physical outcomes that may be linked to each corridor
type are also considered, including: construction disruption, ground level severance and
connectivity, future modal interchange opportunities, creation of new public open space,
transit-oriented development opportunities and value capture.
Rail infrastructure investment by governments and the private sector at the corridor-level
scale has the potential to be city-shaping and any assessment of such projects must include
the broader costs and benefits in terms of urban planning and design. A major conclusion is
that many factors beyond physical construction costs alone must be assessed in order to
provide a genuine analysis of the value of corridor scale rail infrastructure investment.
1. Introduction
After over a decade of relative dormancy, elevated railways are returning to the railway
constructor’s toolbox. The 8.5 kilometre-long elevated section of Brisbane’s ‘AirTrain’ has
served the airport’s two terminals since 2001, meanwhile Sydney’s 4-kilometre-long elevated
‘SkyTrain’ section of the North West Metro is nearing completion. In January 2016,
Melbourne added ‘Sky Rail’ to the Australian lexicon for elevated railways, bestowed upon
the Caulfield to Dandenong Level Crossing Removal Project by a journalist (Johnston 2016).
Almost a month elapsed between the initial media frenzy and the formal government
announcement, accompanied by high-quality images of the proposed concept designs. The
proposal involved 8.2 kilometres of elevated rail in three sections to remove nine level
crossings on Melbourne’s busiest rail corridor. In short order, ‘Sky Rail’ became increasingly
controversial, with much debate on the merits of elevated rail in Melbourne’s established
1ATRF 2016 Proceedings
middle suburbs. In this paper, the authors seek to contribute to more informed discussion on
the value of elevated rail corridors in the Australasian urban environment through
examination of Melbourne’s use of elevated rail in its grade separation program.
This paper is divided into three main sections. Firstly, the Victorian Government’s
methodology for assessing and analysing the costs and benefits of grade separations will be
explored. Secondly, the benefits of elevated rail construction compared to alternatives will be
examined. Thirdly, indicative construction cost profiles will be developed to provide a cost
framework for different types of grade separations and railway to benchmark capital costs of
elevated rail against alternative alignment options.
2. Literature review: institutional biases in benefit
assessment and analysis in Melbourne
Much of Melbourne’s elevated rail debate has concerned the decision-making process
driving methods of grade separation. Many international cities have developed transit
networks involving substantial elevated rail corridors from whose experience much can be
learned. Applying such learning is a fast-evolving area of practice and research in Australia
(see De Gruyter & Currie 2016; Currie 2016; Macdonald 2016; Woodcock & Stone 2016 &
2015; Woodcock 2016; Woodcock & Wollan 2013). The literature is important because of
the role it has played in setting policy and the evidence it provides about institutional thinking
and decision-making.
Towards the end of the 2000s, the removal of level crossings became an increasingly
pressing issue and the Victorian Department of Transport and VicRoads collaborated on
developing a methodology for prioritising grade separation of the 177 level crossings on
Melbourne’s passenger rail system (Taylor & Crawford 2009, Crawford 2010). The Multi-
Criteria Assessment process focused on four main areas: project costs, social aspects,
environmental factors and alignment with strategic objectives for the road network. Project
costs and strategic ‘fit’ were calculated numerically, along with estimates of greenhouse gas
emissions to cover many environmental parameters.
A benefit cost analysis was conducted to weigh up costs of project implementation against
benefits calculated in terms of travel time savings and accident reduction (Taylor & Crawford
2009, Crawford 2010, VicRoads 2014). Social impacts were assessed qualitatively for a
range of criteria, including community severance, visual and noise amenity, development
opportunities, transport connectivity / access, and places of social significance. Impacts on
the local natural environment were generally seen to apply outside urban areas.
Several weightings were trialled to ascertain the effect of different factors across all four
domains, but these could produce contrary results. For example, a weighting towards
environmental factors could lead to local roads being prioritised over arterial roads
(VicRoads 2014:14). Other aspects, such as local accessibility for pedestrians and cyclists
within an area or to public transport were seen as of little importance and not part of the
assessment of ‘transport’ benefits:
Other factors included in the multi-criteria analysis such as amenity
impact were largely matters to satisfy in devising schemes rather than in
prioritising. Similarly, factors such as reducing community severance had
only marginal impact – there needs to be a major transport benefit for a
crossing to be a priority. (Crawford 2010:722)
Ultimately, weightings had to suit strategic road network objectives, with the aim being to
find which level crossings needed to be removed to “maximise the efficiency of the road
network” (VicRoads 2014: 13). This ultimately meant qualitative aspects, even those which
clearly can have hard quantitative measures made of them such as community severance,
development opportunities, or public transport access and connectivity parameters were
disregarded in the prioritisation process.
2Of Skyrails and Skytrains - Elevated rail in the Australasian urban transport environment
Another major framing constraint was seeing each level crossing in isolation. As VicRoads
explained, the assessment was:
...predominantly focused on individual sites and wholly from a road and
rail operations perspective. It did not consider broader network
approaches and wider productivity and economic benefits from a corridor
or strategic perspective. The initial prioritisation was also influenced by
initial costs and benefit cost ratios, which biases the results. While the
multi-criteria assessment was a useful tool, it is not the sole input into
prioritising level crossings for grade separation. (VicRoads 2014: 13)
The problematic nature of the exercise becomes even more apparent when it is realised the
benefit cost calculations were based on cost estimates for grade separation approaches that
lacked an evidence base for their assumptions. For example, under the social criteria of
‘community severance’, assessments incorporated presumptions of the most likely type of
grade separation and its assumed effects on ‘severance’. In this, rail-under-road grade
separations were ranked highest because local road accessibility was improved after
removal of the level crossing; whereas rail-over-road was treated the same as road-over-rail
based on the negative effects at many places where 1960s and ‘70s-vintage road
overpasses resulted in increased severance for pedestrians and cyclists and loss of
economic activity. Despite these documents drawing on evidence from places in Melbourne
where road overpasses have had these effects, there was a failure to also observe the
success of the many extant examples of elevated rail in Melbourne (Woodcock & Stone
2015, 2016; Woodcock 2016). As a result, elevated rail was ranked poorly alongside road
overpasses.
In the 2014 VicRoads report there was an awareness that level crossings were a major
impediment to achievement of the 30-year goals of the Public Transport Victoria Network
Development Plan for passenger rail by “increasing services and capacity of the rail
network” (VicRoads 2014: 6). Indeed, while PTV’s plan indicated dramatic increases in
service frequencies, with some lines tripling or even quadrupling the number of services per
hour in the peak (PTV 2012), VicRoads stated that “if no action is taken to remove the level
crossings, the capacity to be able to run additional train services and provide a higher level
of services to commuters in the future is limited.” (VicRoads 2014: 16).
If a network perspective is taken, the aim of freeing up capacity to allow as many rail
services to run as possible should fit with maximising efficiency of the road network so that
more buses and trams can also be run to feed the enhanced rail network. Taking a
genuinely integrated public transport network focus for Melbourne would make the 116
crossings used by buses and the 3 used by trams priorities for removal (Woodcock & Stone
2015). In their synthesis of the international literature on rail-road crossings, De Gruyter and
Currie (2016) identified the effect of level crossing removals on improved train service
frequencies as one of the major research gaps.
Further implicated in the assessment methodology informing policy in Melbourne (until mid-
2014 at least) is the limited understanding of a connection between transport, especially rail-
based public transport and urban renewal and intensification. Again, the preference for rail-
under-road grade separations is based on an assumption that development opportunities will
exist on top of or adjacent to railway stations. The problem that arises with this is that
because building over the top of railway stations is complex and expensive, VicRoads finds
“there is a limited opportunity to realise positive financial return on the development of rail
land made available by level crossing projects” (VicRoads 2014:12).
This view becomes significant when the potential for integrating land use and public
transport through massive public investment in grade separations are disconnected from
value capture opportunities. As a result, VicRoads reported that “Value capture and
development opportunities have not been considered in the prioritisation of level crossings
3ATRF 2016 Proceedings
as the benefits are generally related to sites that have a very high land value” (VicRoads
2014: 12). As a result of the assumptions in the framework for grade separation assessment,
there is a disconnect between large numbers of level crossings prioritised for removal and
the areas identified for urban renewal in the 2013 Plan Melbourne metropolitan strategy and
for transit-oriented intensification in its predecessor Melbourne 2030. This disconnect would
appear to arise from this assessment framework, or one based on similar assumptions.
Research, and international experience shows, two important benefits of elevated rail run
counter to these assumptions: firstly, elevated rail releases land below the railway that can
be used for a very wide variety of land uses (Woodcock & Stone 2016, 2015; Dovey &
Woodcock 2015; Woodcock 2016; Woodcock & Wollan 2013); and secondly, freeing up this
land can create many kilometres of new frontage, adding significant value to what would
otherwise be large quantities of real estate with back boundaries onto rail corridors.
3. The benefits of elevated rail
At this point, the question raised is: ‘what difference would be made to multi-criteria
assessment analyses by greater understanding of the benefits of elevated rail?’ Part of our
interest is inevitably in the claimed cost differences between types of grade separation, since
these are the major cost component being assessed against the range of benefits that may
accrue from a grade separation project. The cost analysis is dealt with later section 4. In this
section, we will briefly outline the benefits of elevated rail using the same set of criteria as
were used above for social and environmental impacts by VicRoads.
3.1 Safety
The standard numerical assessment of crash statistics and the savings related to traffic and
pedestrian accidents assume they are completely removed by grade separations. At this
level of analysis, elevated rail should perform as well as any other type of right-of-way.
3.2 Community Severance
In addition to freeing up movement of vehicular and pedestrian traffic on roads where level
crossings have been removed, elevated rail releases land in the rail reserve formerly
occupied by the railway tracks. Apart from those parts at either end of the railway ramps
where the structural headroom is too low, the corridor land beneath the viaduct allows free
movement across the former surface railway corridor. Compared to all other grade
separation options except tunnelling (either deep tunnelling or cut-and-cover), elevated rail
maximises reduction of community severance. Where viaducts run for extended lengths,
connectivity gains are maximised, directly reconnecting previously separated communities. If
transport benefits are extended to include accessibility gains for active transport modes
(pedestrians and cyclists) within a local area and to public transport nodes, reduction in
severance needs to viewed with much greater significance than previously. Many local
journeys that may have used cars due to severance could potentially be replaced with active
transport. The other main types of grade separation used in suburban settings will not allow
this to occur.
3.3 Visual Amenity
Visual amenity has been a major issue in Melbourne since January 2016 with the unveiling
of the so-called ‘SkyRail’ project between Caulfield and Dandenong. This has been a
primary issue for resident protesters on parts of the corridor who have actively fought to
change the design to trenched rail. Amenity is a vague term and encompasses a range of
meanings. The primary cluster of issues for residents have related to the visual appearance
of the viaduct, graffiti, and property values. Alongside these concerns, we discuss visual
amenity for the tens of thousands of rail passengers using the line every weekday.
4Of Skyrails and Skytrains - Elevated rail in the Australasian urban transport environment
Lowering rail lines into trenches that pass beneath roads makes the railway disappear from
many angles of view. This is seen as improving visual amenity for those who dislike looking
at rail infrastructure. However, trenches generally do not improve the view for rail
passengers. Contemporary construction standards require trenches to be lined with concrete
retaining walls obviating vegetated cuttings. As has occurred with recently completed
trenches in Melbourne, graffiti artists are quick to add their touch to these extensive blank
canvasses, where their work will mainly be seen by rail passengers. At the same time, the
extensive graffiti on back fences lining the rail reserves remains, along with the mostly
constrained space for landscaping. Protesters have suggested that vandals will find ways to
graffiti viaducts seven to twelve metres above ground and have fed a sensation-hungry
media with ‘mashups’ depicting this. Yet, evidence from elevated roads in Melbourne and
viaducts around the world would suggest it is very difficult to paint graffiti on viaducts.
Contemporary safety standards require trees to be planted sufficiently far away to obviate
them falling across the railway line. Mature trees close to the line must be removed for
construction and replacements are unlikely with average width reserves. On the other hand,
while trees may need to be removed to build rail viaducts, because of the elevation, much
more planting closer to the line is possible to enable screening effects.
Elevated rail dramatically improves visual amenity for railway passengers and the general
travel experience through access to natural light and views. The profile of public transport is
raised within the landscape, enhancing wayfinding and access both to and from stations,
and the bike paths that run beneath the viaducts.
The issue of visual amenity has also been claimed by protesters to result in loss of value for
property directly facing a rail reserve with elevated rail above it. There appears to be no
definitive data about this other than the opinions of real estate agents reported in the media
(Zhou 2015, 2016; Zhou & Robb 2016). Real estate values for properties abutting rail
corridors as they exist can sometimes be lower, but the impacts of reclaiming the rail reserve
may have the opposite effect. Direct access to linear parks, cycling and walking paths more
closely connected to stations and shopping precincts could make these properties more
attractive, especially to developers of medium density housing. Some of the land use zoning
along the Caulfield to Dandenong corridor anticipates intensification and there are already
smatterings of apartment developments that would benefit from more public open space.
Finally, there are a number of properties that will be overshadowed by viaducts and the
Victorian government has offered compensation for these.
3.4 Noise Amenity
In Melbourne at least, at-grade railways are fairly noisy much of the time because of a lack
of maintenance, while level crossings have bells and produce wheel noise from road traffic,
along with requiring train horns to be sounded for safety. New elevated track will produce far
less noise from train wheels, and the noise associated with level crossings is removed.
Sound attenuating barriers can reduce noise transmission further. Whether elevated rail is
quieter than trenched rail may be a moot point, there is evidence that both can produce
unwanted noise effects, however it will be less noisy than rail at-grade.
3.5 Development Opportunities
Decking over railway trenches to provide development opportunities is expensive and
economically viable only when the surrounding land value is the same or greater than the
cost of decking. Developers would find it more profitable to purchase nearby land to
maximise the proximity benefits to the railway and evidence suggests that while property
values are generally higher close to railway stations (LUTI & Mecone 2016), for the
residential market in Melbourne at least, the ‘sweet spot’ may be a block or two back (Zhou
2015, 2016; Zhou & Robb 2016). Furthermore, to make development over a railway trench
viable, greater building height would likely be necessary than adjacent areas, producing
5ATRF 2016 Proceedings
resistance from existing residents. Given enormous costs of grade separation projects, this
explains why these kind of ‘air rights’ developments are viable only in very high land value
areas such as Melbourne’s CBD or certain inner city corridors. To date, most development
on railway land in Melbourne (where it has occurred at all) has been adjacent to stations, not
on top of them. As a case in point, an ‘up to’ 13-storey development has been proposed by
the Victorian government on top of a deck built as part of the recent trenched grade
separation at Ormond on the Frankston Line. However, there has so far been no visual
information or any detail about how this proposal will proceed through the planning process.
By contrast, elevated rail releases land in the rail reserve beneath the tracks and on either
side of the station. This enables a wide variety of land uses to incrementally develop over
time, at a scale in keeping with the area’s pace of development more generally. This allows
for integration of complementary land uses (such as retail, commercial, community and
recreation) within stations and enables them to be better integrated socially with their urban
precincts. This also allows for creation of new public open spaces, whether as station
forecourts or linear parks extending along the railway corridor. In the long run, if activity
centres around elevated stations develop such that high levels of development become
warranted, then elevated rail is no barrier. There are many examples that could be cited
internationally, but Chatswood in Sydney is a good example of how development intensity is
not constrained by elevated rail.
By contrast, the risk is that the new public spaces created beneath viaducts may not be
adequately activated or subject to passive surveillance and so could acquire negative
perceptions. To ensure an actively used and positively viewed public realm requires a
mixture of pro-active place making, a sense of ownership by local community groups,
appropriate planning controls on adjacent land, high quality urban and landscape design,
night-time lighting and good connections to local pedestrian and cycling networks: in short,
good integration of transport and land use planning with high quality spatial thinking. In the
case of the Caulfield to Dandenong ‘SkyRail’ project, a $15 million budget for the undercroft
public realm is being overseen by a trust comprising eminent landscape and other design
and planning professionals.
3.6 Transport connectivity and access
The issue of transport connectivity and access warrants close consideration as it has
significant implications for grade separations. If public transport network efficiency is seen as
a priority to the same extent as the road network, grade separations play a key role in both.
Many transport planners argue the most cost-effective way to improve access to public
transport in dispersed cities like Melbourne is to rationalise bus routes so they form a more
effective network, running them more frequently to form seamless connections with the
heavy rail system. Doing this requires road network efficiency to be maximised and buses
moving freely upon it without being delayed by congestion or level crossings. Level crossing
removal is necessary to make bus (and tram) priority workable. To make the most of free
flowing and frequent buses serving increased capacity passenger rail, grade separation
types should be used that maximise opportunities for inter-modal transfers (Woodcock &
Stone 2016, 2015). Elevated rail provides the greatest flexibility in terms of planning for
current and future interchange arrangements. The degree to which this potential can be
realised depends on maximising access to railway stations by providing as many entrances
as possible to connect them to buses, trams and local pedestrian networks, rather than
single points of entry. To enable this, a culture change is needed here in Melbourne away
from revenue protection-based controlled access, to a European style access culture with
ticket validation technology dispersed among many entry points.
However, the issue of connectivity needs to strategically prioritise grade separations so that
not only is road network efficiency maximised, but the ability of the rail system to run as
many train services as required for a ‘turn up and go’ service is also maximised. Typically,
railways operate as corridors and the spatial distribution of bus routes and railway lines
6Of Skyrails and Skytrains - Elevated rail in the Australasian urban transport environment
means that grade separations need to be planned to enable entire rail corridors to operate
as efficiently as possible, not just the roads that cross them. The benefits of elevated rail are
such that as more of a corridor is elevated, the more overall ground level connectivity can be
achieved.
3.7 Places of social significance
So far, this criterion has been intended to refer to protecting and retaining places that are
valued, such as heritage listed structures like stations, signal boxes and local reserves etc.
Trenched rail projects tend to mean extensive demolition to make way for new structures
and tend to take up more of the rail reserve while exacerbating severance either side of
where level crossings were removed. If carefully planned, elevated rail can minimise this
kind of damage to valued local places. Viaducts can be built over live rail following existing
alignments, and heritage structures may be more likely to be retained. However, beyond
protecting existing places of social significance, we should ask to what extent grade
separations can create new places the community will actively use and value? Here,
elevated rail has significant advantages because space created below the viaduct that acts
to connect previously disconnected paths, streets, and public open spaces along the
corridor. This opens possibilities for creation of many new local places of social significance.
A number of projects internationally propose retrofitting elevated railways to enhance the
potential of ground level connectivity for active transport, such as the Radbahn proposal to
create a continuous 9-kilometre bike path beneath Berlin’s oldest metro, the U1 line
(AllesGerman 2015), and the 16-kilometre Underline project in Miami (The Underline 2016).
Other cities with elevated rail, such as Vancouver, Toronto, New York and Singapore have
used design competitions to create proposals for retrofitting undercroft spaces below railway
viaducts that not only create walking and cycling connections, but add other land uses such
as business and arts incubators, markets, various recreational uses as well as potential to
act as ecological corridors for urban flora and fauna. We see here a shift in attitudes towards
the value of space created as a by-product of major transit infrastructure built at a time when
cities functioned very differently. As local demographics change and economies shift from
manufacturing to services, neighbourhoods with elevated rail can change to serve the needs
and aspirations of the people who live there and pass through them. In a city like Melbourne,
new elevated rail can be designed with such uses in mind to ensure that not only does
capital investment free up transport movements, but is also a catalyst for place-making.
3.8 Local Natural Environment
In a similar vein to above, the impact in local natural environments was assessed primarily in
terms of damage minimisation, hence why it was felt to apply mainly to grade separations
outside urban areas. Again, we ask to what extent can rail infrastructure projects create new
local natural environments that provide extensions to local public open spaces, wildlife
habitats and connected ecological corridors? Here elevated rail clearly can make a
potentially radical and transformative contribution. An analysis elsewhere by one of the
authors (Woodcock 2016) indicates that if grade separations in Melbourne where prioritised
on the basis of enhancing public transport network effects, then at least 146 kilometres of
railway corridors, involving the redevelopment of 96 stations would be required. If this were
implemented as elevated rail (wherever appropriate from a rail operations viewpoint), then
this would create a significant transformation in the public open space network, with over
400 hectares of new linear parks. In the case of the Caulfield to Dandenong ‘SkyRail’
project, 22 hectares of new public open space will be reclaimed from the rail reserve,
allowing the re-connection of a series of local public open spaces, improving ecological
connectivity. These impacts on local natural environments within urbanised areas are much
in need of further research.
7ATRF 2016 Proceedings
4. Recent grade separations in Australia and New Zealand
Much recent debate over elevated railways in Australia has stemmed from a major program
of grade separations on the metropolitan rail network of Melbourne, Australia that has taken
place in recent years. The program largely aims to improve efficiency of the metropolitan
road network and manage increased congestion, with increased rail network capacity as a
secondary, but still important priority (VicRoads, 2014: 6). The size and scope of the
Victorian Government’s target to remove 50 level crossings by 2022 has necessitated
thinking on engineering and urban design for level crossing removals to shift from ‘spot’
treatments of particular level crossings to ‘corridor’ level treatments of multiple level
crossings along a suburban railway line.
While Sydney largely grade separated its rail network from upgrading its arterial road
network from the 1950s to the 1990s, grade separations on Melbourne’s rail network were
much more sporadic and not prioritised by the road and rail agencies to the same extent
(Martin, 2012b). While not faced with the same levels of road congestion as Sydney and
Melbourne, other Australian capital cities, particularly Brisbane and Perth are investing in
grade separations on their rail networks in response to growing populations and road
network congestion, particularly for freight transport (Infrastructure Australia, 2015: 2).
Not all grade separations in Australian and New Zealand cities have been undertaken purely
for transport network efficiency reasons. Urban renewal in either CBD fringe or suburban
town centre sites have been cited as important factors in some grade separations,
particularly in Auckland (New Lynn) and Perth (Citylink).
Another important factor driving grade separations is the need to improve rail network
efficiency by development of ‘flying junctions’ on rail networks to increase capacity and
reduce conflict between trains on ‘at grade’ junctions. Development of flying junctions has
been particularly important in cities with significant interaction between passenger and
freight trains such as Adelaide, Newcastle and Sydney and increasing capacity on busy rail
corridors where ‘at grade’ junctions exist.
The development of a dataset of grade separations was deemed as an important research
outcome of this project to better guide decision makers on potential costs and benefits of
different types of grade separations undertaken on Australia’s urban rail networks.
4.1. Developing a capital cost dataset for grade separation
As part of this paper’s data collection process, a capital cost dataset on recent grade
separations on Australia and New Zealand rail networks was developed.
This builds upon previous work on capital costs for public transport infrastructure projects in
Australia and New Zealand by the secondary author (Martin, 2011 & 2012a). Using the
methodology outlined in these works, capital costs in dollars of the day were escalated into
2016 dollars using Producer Price Index data (ABS, 2016 & Statistics NZ, 2016), while
conversion from $NZ to $A was achieved using the Reserve Bank of Australia’s long-term
currency conversion tables (RBA, 2016).
Alongside the three typologies of grade separations previously identified by Woodcock &
Stone (2016: 24) of ‘rail over road’, ‘rail under road’ and ‘road over rail’, a fourth of ‘rail over
rail’ is used in this study to capture five significant ‘flying junction’ grade separation projects.
From this investigation, a dataset of 26 grade separation projects were identified on
Australian and New Zealand urban rail networks between 1996 and 2016 and are ranked in
chronological order in Table 1. This table outlines locations, types of grade separation,
project costs, number of grade separations undertaken, the length of new road or rail
constructed in each project and whether a new station was constructed.
8Of Skyrails and Skytrains - Elevated rail in the Australasian urban transport environment
Not all grade separations were able to be identified in the dataset, as either many grade
separations were built as part of larger road or rail upgrading projects, or where costs for
grade separation works could not be disaggregated from overall project costs. Based on this
dataset, the quantum of spending on identifiable grade separations over the past 20 years is
approximately $2.9 billion in 2016 Australian dollars.
4.2. What the dataset tells us
As indicated in Table 1, a broad range of project costs were observed for grade separations
in Australian and New Zealand cities. Generally, ‘Road over Rail’ grade separations were
the lowest cost projects, with nine projects identified in a capital cost range between $23
million and $99 million, depending on the complexity of the grade separation and supporting
works such as utilities relocation, road duplication or intersection upgrading. It is likely that
the low costs for this type of grade separation project is based on the way these projects
interface with the railway (often at intermediate locations between stations), the location of
road/rail overpasses (middle and outer suburbs) and the lack of new or rebuilt railway
stations kept costs low compared to alternative options.
‘Rail under Road’ grade separations tend to be more expensive, based on a sample size of
10 projects, ranging in capital cost between $51 million and $169 million and an outlier
project (Perth Citylink) costing $364 million. The three rail-under-road grade separations built
in the decade 1998-2007 offered significantly lower costs (from $43 to $86 million) compared
to the seven post-2009 projects ( from $131 to $364 million). While more research needs to
be undertaken into the escalation in costs from pre-2007 and post-2009 rail-under-road
grade separations, some preliminary thoughts on the causes could involve: complexity of the
project, constraints of the worksite, the need to construct new stations and intermodal
transfer facilities in the project scope and the construction of long trenches and approach
ramps to transition the vertical alignments of the rail corridor in and out of trenches.
Increasingly, Rail under Road projects also demand construction of new railway stations,
modal interchanges and park and ride facilities, particularly in Melbourne where level
crossing removal sites are often adjacent to railway stations. Such new public transport
facilities trigger extensive additional redevelopment works to build new public transport
infrastructure compliant with modern accessibility requirements and provide urban design
outcomes that meet community expectations on accessibility and placemaking, particularly
those stations in suburban town centres. The Victorian Government has recently
acknowledged the ‘additional works’ required to improve public transport nodal infrastructure
(stations, interchanges) along with enhanced urban design and amenity as part of grade
separation projects has added approximately $1 billion dollars to the overall cost of its level
crossing removal program (Willingham, 2016).
The ‘outlier’ Perth Citylink project was an exceptional project, taking place on the edge of the
Perth CBD and involved sinking the Fremantle rail line, extending the Joondalup line tunnel,
rebuilding the interfaces between the Fremantle and Joondalup lines and extensive
upgrading, platform rearrangement and access improvements to Perth railway station. The
$364 million cost of this project was a small component of a wider $1.3 billion urban renewal
project to sink the railway below the surface, construct a deck over the top of the railway and
redevelop 13.5 hectares of land on the western edge of Perth’s city centre (MRA, n.d.).
The five Rail-under-Rail grade separations ranged between $46 million and $265 million,
with three being relatively high cost projects (at over $100 million each). All of these projects
were built within the rail corridor and fall into three essential categories: improving separation
between passenger from freight traffic on rail networks on suburban rail networks (at
Flemington Junction, Goodwood Junction and North Strathfield Junction), the separation of
the high capacity bulk coal railway from the interstate rail network (Sandgate Flyover) or
provide a new, suburban rail route with a grade separated junction (Y-Link). The outlier
project (North Strathfield Rail Underpass) cost over $265 million and involved ramping of the
9ATRF 2016 Proceedings
Table 1: Road/Rail and Rail/Rail grade separations in Australia & New Zealand 1996-2016
Year Location Type City Cost $ Cost $ Grade Length New
OTD 2016 Seps (km) Station
1996 Merrylands-Harris Rail Under Rail Sydney $80.0 $150.5 1 1.6 -
Park ‘Y-Link’
1998 Subiaco, Perth Rail Under Road Perth $35.0 $64.8 2 0.9 1
1998 Boronia & Dorset Rds Rail Under Road Melbourne $28.0 $51.4 2 1.1 1
Boronia
1999 Flemington Junction Rail Under Rail Sydney $31.0 $57.4 1 0.6 -
2001 Westall Rd, Clayton Road Over Rail Melbourne $37.0 $63.3 1 0.8 -
2006 Sandgate flyover Rail Over Rail Newcastle $80.0 $106.1 1 0.9 -
2007 Taylors Rd, Keilor Road Over Rail Melbourne $54.0 $68.3 1 1.1 -
Downs
2007 Middleborough Rd, Rail Under Road Melbourne $66.0 $86.3 2 1.3 1
Laburnum
2007 Somerton Rd, Road Over Rail Melbourne $34.0 $39.6 1 0.7 -
Roxburgh Park
2008 Kororoit Rd, Altona Road Over Rail Melbourne $48.5 $55.2 1 1.1 -
2009 Dynon-Port rail link Road Over Rail Melbourne $116.0 $136.8 3 - -
2009 Beaudesert Rd, Acacia Rail Under Road Brisbane $114.0 $134.8 1 1.4 -
Ridge
2009 Mawhinney St, Road Over Rail Brisbane $70.0 $82.1 1 1.2 -
Beerwah
2010 Daddow Rd, Kewdale Road Over Rail Perth $19.8 $22.8 1 1.0 -
2010 New Lynn, Auckland Rail Under Road Auckland $114.1 $168.2 2 1.0 1
2010 Springvale Rd, Rail Under Road Melbourne $140.0 $162.4 1 0.8 1
Nunawading
2010 South Road Tram Rail Over Road Adelaide $30.0 $34.8 1 0.6 1
Overpass
2013 CityLink, Perth Rail Under Road Perth $360.0 $364.4 0 1.3 -
2013 Goodwood Junction Rail Under Rail Adelaide $45.0 $46.0 1 0.6 -
2014 Robinson Rd, Road Over Rail Brisbane $98.0 $98.9 1 1.0 -
Geebung
2014 Telegraph Rd, Bracken Road Over Rail Brisbane $80.4 $80.5 1 0.8 -
Ridge
2014 Mitcham & Rooks Rd, Rail Under Road Melbourne $192.0 $194.0 2 1.3 1
Mitcham
2014 Springvale Rd, Rail Under Road Melbourne $159.0 $159.9 1 1.0 1
Springvale
2015 Lloyd St, Midland Rail Over Road Perth $80.0 $79.6 1 0.6 -
2015 North Strathfield Rail Under Rail Sydney $264.0 $265.2 1 4.0 1
Junction
2016 Burke Rd, Glen Iris Rail Under Road Melbourne $131.0 $131.0 1 0.8 1
rail corridor, a short section of driven tunnel, extensive new trackage and station
reconstruction. As with Rail-under-Road projects, the range of costs is dependent upon each
project’s level of complexity, level of disruption and the amount of additional work for new or
upgraded stations and track incorporated into the project scope.
10Of Skyrails and Skytrains - Elevated rail in the Australasian urban transport environment
Only two examples of ‘Rail over Road’ grade separations were identified in this process,
costing between $34 million (South Road Tram Overpass) and $80 million (Lloyd St,
Midland). The relatively low cost of these projects compared to other types of projects may
be due to the engineering differences between light and heavy rail (South Road), their
relatively easy constructability in existing corridors and reduced disruption to both road and
rail traffic through shorter road and rail network closures and alternative road traffic routes.
A scatter plot showing the capital costs (in 2016 $ million) of each grade separation category
in chronological order is shown in Table 2 below.
Table 2: Capital costs of grade separations by type in Australia and New Zealand 1996 - 2016
$400
Rail Under Road
$350 Rail Under Rail
$300 Road Over Rail
Rail Over Road
$250
$200
$150
$100
$50
$0
0 2 4 6 8 10
From Table 2, it is clear that there are a band of 12 relatively ‘simple’ grade separation
projects (almost half the sample) across all formats in the sub-$100 million cost range,
particularly Road-over-Rail, Rail-over-Road and less complex Rail-under-Rail projects.
Table 3 below shows the indicative range of costs for each of the four types of grade
separations based on the Table 1 dataset, with the median cost for each type shown. For
the purposes of road/rail grade separations, Table 3 clearly shows the lower average cost
and lower range of costs for the Road over Rail or Rail over Road (elevated rail) alternative
options compared to Rail under Road (entrenched rail) grade separation options.
Table 3: Typical capital cost range and median cost of grade separations in Australia and New
Zealand 1996 - 2016
$400
$350
$300
$250
$200
$150 $153.6
$125.0
$100
$78.2
$50 $57.2
$0
Rail Under Road Rail Under Rail Road Over Rail Rail Over Road
11ATRF 2016 Proceedings
Table 4: Indicative capital cost data for elevated, trenched and tunnelled rail
Cost Cost Cost/km
Length
Project Name City R-O-W % Elevated Date Stations ($A (2016 (2016
(km)
OTD) $A PPI) $A PPI)
Elevated
Airtrain BNE 100% elevated 2001 8.5 2 $220.0 $375.6 $44.2
railway
Sydney Metro NW Elevated
SYD 100% elevated 2016 4.2 0* $389.7 $394.5 $93.9
Skytrain railway
Rowville Line Elevated 90% elevated /
MEL 2012 5.8 0* $290.0 $301.8 $52.0
(Estimated) railway 10% trenched
Entrenched
New Lynn trench AKL 100% trenched 2010 1.0 1 $106.2 $119.4 $119.4
railway
Beaudesert Rd Acacia Entrenched
BNE 100% trenched 2009 1.4 0 $114.0 $134.8 $96.3
Ridge Grade separation railway
Perth CityLink Tunnelled 100%
PER 2014 1.3 0 $335.0 $339.1 $260.8
Fremantle line tunnel railway tunnelled
Elevated 7% tunnelled /
Skytrain Expo Stage 1 YVR 1986 21.4 15 $917.8 $2,248.4 $105.1
railway 93% elevated
Elevated
Skytrain Expo Stage 2 YVR 100% elevated 1990 3.1 3 $196.3 $369.4 $119.2
railway
Elevated
Skytrain Expo Stage 3 YVR 100% elevated 1994 4.4 2 $149.5 $288.6 $65.6
railway
5% tunnel / 6%
Elevated
Skytrain Millenium Line YVR trench / 89% 2002 20.5 11 $1,367.0 $2,136.2 $104.2
railway
elevated
Elevated 48% tunnelled
Skytrain Canada Line YVR 2009 19.2 16 $1,671.5 $1,827.5 $95.2
railway / 52% elevated
Elevated 82% elevated /
Skytrain Evergreen Line YVR 2013 11.0 6 $889.0 $910.6 $82.78
railway 18% tunnelled
South Morang rail Entrenched
MEL 100% trenched 2012 3.5 2 $261.0 $274.8 $78.5
extension railway
Epping-Chatswood Tunnelled 100%
SYD 2009 12.5 5 $2,350.0 $2,770.5 $221.6
Railway Line railway tunnelled
Sydney Airport Rail Tunnelled 100%
SYD 2000 10.0 0* $762.0 $1,335.1 $133.5
Link railway tunnelled
Perth-Mandurah Tunnelled 100%
PER 2007 2.2 2 $398.1 $520.8 $236.7
Railway 'Package F' railway tunnelled
Sydney Metro NW Tunnelled 100%
SYD 2016 15.5 5* $1,150.0 $1,183.0 $76.3
Tunnel contract railway tunnelled
Mitcham & Rooks Road Entrenched
MEL 100% trenched 2014 1.6 1 $192.0 $195.3 $122.0
Grade separation railway
Elevated
Bangkok Skytrain BKK 100% elevated 1999 23.1 23 $1,038.8 $1,742.9 $75.4
railway
Alameda mid-corridor Entrenched
LAX 100% trenched 2002 16.0 0 $518.4 $905.9 $56.6
trench railway
43% elevated /
Caulfield – Dandenong Elevated
MEL 5% trenched / 2018* 19.4 5 $1,600.0 $1,600.0 $82.5
corridor (Estimated.) railway
52% surface
12Of Skyrails and Skytrains - Elevated rail in the Australasian urban transport environment
5. Elevated, trenched and tunnelled rail
This next section deals with capital cost data for elevated, trenched and tunnelled railway
lines, as there has been relatively little research into the capital cost differences of all options
in an Australasian context. Some international knowledge of the capital cost differences for
each approach exists, particularly the work of Flyvbjerg, Bruzelius & Van Wee (2008) on
capital costs for international urban rail projects, but generally knowledge on capital costs in
Australasian conditions is relatively thin, apart from the work of this paper’s secondary
author (Martin 2011, 2012a) and work done as part of inquiries for the Productivity
Commission and Parliamentary Committee reports in NSW and Victoria.
5.1. Capital cost data for elevated, trenched & tunnelled rail
The lack of capital cost estimation data for elevated and entrenched rail options in Australia
and New Zealand is a serious gap for policy makers and transport planners. It is telling, that
the key document providing guidance on economic appraisal and modelling of transport
projects (National Guidelines for Transport System Management in Australia) contains only
per-kilometre cost estimates for at grade and tunnelled alignments, with no estimates for
entrenched or elevated right-of-way options in its list of indicative costs (ATC 2006: 43).
A capital cost dataset for elevated, trenched and tunnelled rail construction is shown at
Table 4 below. The small group of actual and proposed Australian rail projects were
augmented with global examples (Vancouver, Bangkok). For trenched rail, international data
(LA’s Alameda trench) was augmented with data from grade separations in Australia and
New Zealand cities where trench lengths longer than one kilometre were constructed.
Based on the capital cost estimates from Table 4 above, indicative per-kilometre capital cost
estimates for each construction type (at-grade, elevated, entrenched and tunnelled) of heavy
rail construction were estimated and shown in Table 5 below. Estimates of the capital costs
of at-grade and tunnelled rail use median indicative costs quoted by the Australian Transport
Council (ATC) escalated to 2016 dollars and sensitivity tested against the secondary
author’s dataset of actual construction costs in 2016 dollars (Martin 2012a). Estimates of
median elevated and tunnelled rail costs in Table 5 use figures developed from Table 4.
In Table 5, indicative capital cost ratios from elevated and entrenched rail show that both
construction types cost approximately twice as much as at-grade, with ratios of 2:1 and 2.2:1
respectively, with tunnelled rail at 4.3:1. Differences in cost ratios between right-of-way types
are close to those cited in Flyvbjerg et al (2008: 25), of elevated rail at 2-2.5:1 and tunnelled
rail at 4-6:1. The surprising discovery in Table 5 was the slightly higher capital cost ratio
(10%) for entrenched rail (2.2:1) over (2:1). The authors offer that higher costs of
infrastructure relocation and transport network disruption encountered in entrenched
construction may account for the 10 percent construction cost premium.
Table 5: Indicative median per-kilometre cost of rail construction types (2016 $A)
At grade rail $43
Elevated rail $86
Entrenched rail $95
Tunnelled rail $186
$0 $20 $40 $60 $80 $100 $120 $140 $160 $180 $200
13ATRF 2016 Proceedings
6. Concluding discussion
In this paper, we have shown that elevated rail has a small per-kilometre capital cost
advantage (around 10%) relative to entrenched rail, based on thorough examination of
capital cost data for rail construction in a range of rights-of-way. The authors also contend
that along with this slight advantage in construction cost, elevated rail has many of the
benefits of the most expensive tunnelling options.
Analysis of projects shows the relative novelty of elevated rail construction in an Australian
context (with only two new-build examples in the past 15 years) means the ability of
transport policy makers to properly evaluate elevated rail in comparison to other methods
has not yet caught up with engineering decisions to proceed on elevated rail construction in
Melbourne and Sydney. Of particular concern is that the current revision of the 2005 ATC
Guidelines does not appear to have taken account of elevated and entrenched rail rights of
way into account for its costing guidelines, with existing arrangements still in use.
In addition, we have yet to see an updated list of level crossing removal priorities for
Melbourne that adequately addresses the policy implications of the recent body of research
in this area. The current list of 50 removals taken to the last Victorian state election has so
far seen one amendment: one additional crossing removal at Cheltenham due to proximity to
the one already mandated for removal. The list of 50 only bears a passing resemblance to
the ALCAM safety rankings, it lacks strong alignment with rankings for high modal
interchange locations (Woodcock & Stone 2015) and is disconnected from many corridors
and places identified for urban renewal. The next 50 grade separations should be prioritised
to create conditions for enhanced tram and bus priority and rail service frequencies.
Elevated rail’s benefits derive from two main sources. Firstly, it is less disruptive to services
and utilities in construction and can be built over a live rail environment, reducing network
disruption. This advantage is where most of the cost savings associated with construction
are claimed to lie. Secondly, elevated rail releases land in the rail corridor formerly occupied
by tracks. The third, and main benefit is the potential for incorporating a wide range of
ground level land uses, including retaining heritage structures and creating new public open
spaces to better integrate stations and the railway corridor into the surrounding natural and
built environment, while enhancing local movement networks and access to public transport.
In these discussions of costs and benefits, it has been shown that the exercise to determine
costs and benefits needs to be recalibrated and incorporate new thinking about how to value
a wide range of benefits going beyond differences in construction costs. An additional aspect
that also needs consideration is the political dimension of focusing solely on costs alone.
The perception that elevated rail is less expensive than other options has allowed opponents
to characterise it as ‘cheap and nasty’. In mature suburban environments of Australian cities
where aesthetic notions of neighbourhood character have driven planning and politics for at
least three decades, this can be a powerful refrain. There is a risk that the benefits of
elevated rail could be foregone if the intestinal fortitude to ensure these benefits are fully
realised is lacking among the political class and the bureaucracy.
The major potential negatives hinge on how well elevated rail is executed, ensuring that
design and construction of the infrastructure is as good as it can possibly be. Stations must
have full-weather protection, universal access, and multiple entrances to paid areas,
alongside the best possible transfer arrangements to other modes to make the most of the
potential offered by elevated rail. Viaducts need to be imaginatively designed as they will
become prominent parts of the urban landscape and the daily lives of people who pass over,
under and around. Open spaces need to be managed by locally engaged and accountable
groups with a strong sense of ownership. Where deemed necessary, noise attenuation
barriers and privacy screens should be opportunities for creativity and individuality. In the
final analysis, elevated rail done well provides the best potential for making the great public
places that the best of public transport infrastructure around the world has fostered.
14Of Skyrails and Skytrains - Elevated rail in the Australasian urban transport environment
References
Australian Bureau of Statistics, 6427.0 – Producer Price Indexes, Australia.
AllesGerman 2015, Berlin Radbahn – A Paradise for Cyclists, 30 November,
http://allesgerman.com/berlin-radbahn-covered-cycling-route/
Australian Transport Council, 2006, National Guidelines for Transport System Management
in Australia (2nd Edition), Volume 4 – Urban Transport, Canberra.
Carey, A 2016 ‘High-rises plans for suburban stations to help fund level crossing removals’,
25 July
Carey, A 2015, ‘Level crossings: VicRoads list of Melbourne’s worst overlooked by Labor’,
The Age, 26 October.
Crawford, R 2010, ‘Prioritising Level Crossings in Melbourne for Grade Separation’,
Proceedings of the Conference on Railway Engineering (CORE) Rail, Rejuvenation and
Renaissance, Wellington, NZ, 12-15 September, pp. 716-722.
Currie, G 2016, Research Perspectives & Comments on Rail Grade Crossing Removal in
Melbourne, Victorian Planning and Environmental Law Association / Ratio seminar: Is
SkyRail a new crossroad for Melbourne transport and planning?, Melbourne, Treasury
Place, 28 April.
De Gruyter, C & Currie, G 2016, ‘Impacts of Rail-Road Crossings: International Synthesis
and Research Gaps’, Transportation Research Board 95th Annual Meeting, January.
Dovey, K. & Woodcock, I (Eds) 2015, Intensifying Melbourne: Transit-Oriented Urban
Design for Resilient Australian Cities - major report on ARC Linkage Project 100200590,
Melbourne School of Design, The University of Melbourne with Monash Architecture, Design
& Art, Monash University.
Flyvbjerg, B, Bruzelius, N & Van Wee, B 2008, ‘Comparison of Capital Costs per Route-
Kilometre in Urban Rail’, European Journal of Transport Infrastructure and Research, Vol. 8
(1), pp. 17-30.
Infrastructure Australia 2015, Australian Infrastructure Audit – Our Infrastructure Challenges:
Executive Summary, Sydney.
Johnston, M 2016, ‘Sky rail for Pakenham Cranbourne outlined in secret Andrews
Government plans, Herald Sun, 11 January
Johnson, S 2016, ‘OVGA gives Melbourne ‘sky rail’ tick of approval’, 29 February,
http://architectureau.com/articles/ovga-gives-melbourne-sky-rail-tick-of-approval/
Lill, C and Kane, T 2012, ‘A Case Study for Engagement with Passionate Groups: Rail
Crossings on the Caulfield to Dandenong Railway Line’, Papers of the 35th Australasian
Transport Research Forum, Perth.
LUTI Consulting & Mecone Planning 2016, Transit and Urban Renewal Value Creation:
Hedonic Price Modelling Assessment of Sydney’s Key Transit and Transit-Oriented Urban
Renewal Investments (2000-2014), CRC for Spatial Information.
Macdonald, A 2016, Local Government Guide to Level Crossing Removals, Metropolitan
Transport Forum, Urban Interface, Melbourne, Online publication:
http://www.mtf.org.au/site/files/ul/data_text12/5144164.pdf
Martin, S 2011, ‘Reviewing the last decade of public transport infrastructure projects in
Australasia’, Papers of the 34th Australasian Transport Research Forum, Adelaide.
15ATRF 2016 Proceedings
Martin, S 2012a, ‘Costing Australian passenger rail projects 2000-2012: How much did we
pay and what did we get?’ Proceedings of the Conference on Railway Engineering (CORE)
Global Perspectives, Brisbane, 10-12 September, pp. 545-558’.
Martin, S 2012b, ‘Victoria’s Level Crossing Removal Program’, Transit Australia, April, pp.
100-105.
Metropolitan Redevelopment Authority, n.d., Perth City Link - fact sheet.
PTV, 2012, Network Development Plan: Metropolitan Rail: Overview, Melbourne.
Reserve Bank of Australia (n.d.), Historical Data – Exchange Rates.
Statistics NZ, Producer Price Index.
Taylor, J & Crawford, R 2009, ‘Prioritising Road-Rail Level Crossings for Grade Separation
Using a Multi-Criteria Approach’, Papers of the 32nd Australasian Transport Research
Forum, Auckland.
The Underline 2016, https://www.theunderline.org/
VicRoads 2014, Strategic Framework for the prioritisation of level crossings in metropolitan
Melbourne, Version: 1.1, 17 June, Melbourne.
Willingham, R 2016, ‘Multi-million-dollar price tag put on level crossing removal plan's hidden
extras’, The Age, 20 September.
Woodcock, I & Stone, J 2016, The Benefits of Level Crossing Removals: Lessons from
Melbourne’s historical experience, RMIT University/University of Melbourne, Melbourne.
Woodcock, I 2016, ‘The Design Speculation Action Research assemblage: “Transit for All”
and the transformation of Melbourne’s passenger rail system’, AECURN Special Issue,
Australian Planner.
Woodcock, I & Stone, J 2015 ‘Grade separations and intermodal transfer at railway stations
in Melbourne’, Proceedings of the 37th Australasian Transport Research Forum, Sydney.
Woodcock, I & Wollan, S 2013, ‘Public Use Zone: A new paradigm for suburban rail station
design for Australian cities’, State of Australian Cities National Conference, Sydney
Zhou, C 2015, ‘Level crossing removals set to give property prices a boost in Melbourne’s
south east’, Domain, 31 May.
Zhou, C 2016, ‘Those closest to sky rail may not be the biggest losers’, Domain, 3 April.
Zhou, C & Robb, K 2016, ‘Melbourne sky rail and railway crossing removal impact property
prices’, Domain, 21 February.
16You can also read
